Read the Original Article at http://www.informationweek.com/news/showArticle.jhtml?articleID=240156499
Amazon Web Services (AWS) is lowering the price of its on-demand Relational Database Service (RDS) by 18% to 28%. The service allows a user to commission an Oracle, MySQL or SQL Server database on EC2 and be charged by the hour for its use.
Pricing on reserved instances under three-year contracts was cut even more. All the new prices took effect June 11.
MySQL and Oracle charges have been reduced 18%. An Oracle instance user must bring license rights from a purchase of Oracle made on premises or purchase them outright for use on Amazon. Oracle on an extra-large server is now 64 cents an hour. MySQL, which is open source code and doesn't have a "bring your own license" requirement, is also 64 cents an hour.
Microsoft's SQL Server has been reduced 28% for the on-demand, bring your own license instance. The price for a standard deployment, extra large server instance, is now 88.8 cents an hour.
[ Want to learn more about how Amazon's low-margin pricing is impacting the industry? See Amazon Cloud Pricing Threatens Tech Titans. ]
All of the above pricing applies to AWS's most heavily trafficked data center complex in Ashburn, Va., known as US East. Prices vary from one Amazon region to another, with US East usually having the lowest prices, as is the case with the database service offerings.
Amazon's chief cloud evangelist Jeff Barr posted a comparison of before and after pricing for reserved instances in a June 3 blog. He did the comparison only for three-year reserved instances, where the price reductions show up in dramatic fashion. To use reserved instances, a customer signs up for a set number of hours per year over a one-year or three-year period and makes an upfront payment. The customer then pays a lower hourly rate than on-demand users.
Reserved instance pricing for MySQL and Oracle for three-year contracts have been reduced 21% in US East and US West (Oregon). MySQL on an M2 or memory optimized, extra-large server before June 1 was priced at $4,441. After June 1, it is priced at $3,507.
By way of contrast among regions, the same database instance in Amazon's Silicon Valley data center, US West (Northern California), was priced at $6,044; after June 1, it is $4,410. The reduction is 27% in those cases.
The same database instance in the Amazon's GovCloud, also located at its US West Oregon data center, used to be $4,835 for a three-year reserved instance contract. After June 1, it moved to $4,217, a 13% reduction.
As usual, the Amazon announcement tended to tout the fact of the reductions without offering any insight into the rationale for why they varied widely among regions, or why reserved instances tend to fare well in most AWS price reduction schemes. The reserved instances appear to provide Amazon with something close to a steady-state customer and make its overall capacity planning problem simpler than with the widely fluctuating, on-demand customers. Amazon's growing emphasis on reserved instances reflects its desire for longer term, more predictable customers.
"We have made a lot of progress since releasing Amazon RDS just 3.5 years ago. ... You have the ability to provision up to 30,000 I/O operations for demanding production workloads, encryption at rest using Oracle's Transparent Data Encryption, and simple disaster recovery using multi-availability zones and read replicas," Barr wrote in the announcement.
The Relational Database Service now has "tens of thousands" of customers, Barr wrote in an earlier blog.